Skip to main content Skip to main navigation



AI-based anonymization in medicine

  • Duration:

The use of artificial intelligence (AI) offers the opportunity to fundamentally support and revolutionize knowledge-intensive activities. In business and many scientific disciplines, the changes brought about by AI are unmistakable. However, the use of AI in medicine is still in its infancy. This can be attributed to the fact that medical data requires special protection and is strictly confidential, which makes it difficult to use this data in research and development projects. In medicine, artificial intelligence methods can be used in everyday treatment or research, for example. Here, decision support systems can show physicians optimal treatment paths as well as their prospects of success. However, the development of such AI solutions requires data from hospital information systems, which are subject to the General Data Protection Regulation (GDPR) and therefore pose significant challenges for data use.

To make medical data more accessible for research and development, methods are used to depersonalize the data. The methods used in this context are anonymization, which describes the removal of the personal reference by modifying the data, and synthesization, which describes the generation of artificial data without a real personal reference, but which thereby maps the statistical properties of the original data. However, these methods are rarely used in medicine because their use is associated with challenges arising from the complexity and heterogeneity of medical data. It is also necessary to reconcile the technical possibilities with the existing legal requirements in order to profitably use the generated data.

The goal of the project is to develop and establish an anonymization platform that combines anonymization and synthesis methods in a hybrid approach to enable the generation of large amounts of medical data. The data should have no personal reference and therefore be easier to use in terms of data protection. This should significantly simplify access to medical data for the development of AI solutions and thus promote research at the interface between AI and medicine.


  • Institut für Medizinische Informatik Westfälische Wilhelms-Universität Münster (WWU) - Hauttumorzentrum der Klinik für Hautkrankheiten Westfälische Wilhelms-Universität Münster (WWU) - AG Medizininformatik – Berlin Institute of Health Charité (BIH) - DATATREE AG

Assoziierte Partner

  • VISUS Health IT GmbH - Westdeutsches Tumorzentrum Essen - Universitätsklinikum Essen - Digital Avantgarde GmbH


BMBF - Federal Ministry of Education and Research


BMBF - Federal Ministry of Education and Research
EU - European Union


EU - European Union

Publications about the project

David Jilg; Joscha Grüger; Tobias Geyer; Ralph Bergmann

In: BPM 2023 Best Dissertation Award, Doctoral Consortium, and Demonstration & Resources Forum. BPM Demo Track (BPMTracks-2023), Ceur-WS, 2023.

To the publication